Hey everyone 👋 Just wrapped up a project I've been building for the past few weeks and wanted to share it here since this community will appreciate the technical side. I built a modular 5-agent AI system for restaurant automation using n8n. Here's the architecture: 🔧 Tech stack: → n8n v2.6 (self-hosted + ngrok) → Supabase (PostgreSQL) — central database → OpenAI GPT-4o-mini + GPT-4o + DALL-E 3 → SerpAPI (Google Maps, TripAdvisor, Yelp scraping) → OpenWeatherMap API → Telegram Bot + Gmail OAuth2 → HTML/JS dashboard on Netlify ⚙️ The 5 agents: 1. Reputation Manager — scrapes reviews from 3 platforms, runs sentiment analysis, upserts to Supabase 2. Reservations Bot — multi-step Telegram conversation flow with session management in Supabase 3. Weather Menu Promoter — fetches weather, matches menu items by weather_tags, sends Telegram recommendations 4. Inventory Manager — reads stock, builds AI prompt, parses JSON response, sends low-stock alerts 5. AI Menu Photographer — dual mode: GPT-4o Vision analysis OR DALL-E 3 generation, stores in Supabase Storage A few things I learned building this: - Merge node must use Append mode (not Choose Branch) to avoid blocking - Telegram Parse Mode should be removed to prevent MarkdownV2 errors - OpenAI response lives at response.output[0].content[0].text Full workflows on GitHub (sanitized, no credentials): github.com/Evgeniy1970-ai/restaurant-ai-suite Happy to answer questions or discuss the architecture. What would you have built differently?

My very first experience creating an AI agent using n8n. I am excited to share the results of my 3-month journey here. Transitioning from a career as a PhD in Archaeology to building AI systems has been a profound experience. Today, I am presenting my fully functional 'International Content Scout'. What you see on the screen: This is a 'Man-in-the-Loop' architecture built in n8n. - The Creative Core: I send a raw thought via Telegram. The AI Agent (GPT-4o) instantly generates expert content, video scripts, and engagement posts in English, Polish, and Spanish. Simultaneously, DALL-E 3 creates a unique visual for the topic. - The Control Center: The system doesn't post anything without my word. I receive everything in Telegram and use Interactive Buttons to Approve or Decline. - The Execution: Once approved, the content is queued for direct publication via Facebook Graph API and LinkedIn API. About the Development: This project was co-architected with Gemini 1.5 Pro, an advanced AI model from Google. During weeks of collaboration, Gemini acted as a mentor, technical consultant, and debug partner, bridging the gap between scientific methodology and modern AI automation. This demonstrates that it is never too late to learn how to find truth in the digital era! Check it out here: https://github.com/Evgeniy1970-ai/ai-content-scout